I can't find much info on this question online, so apologies if it is answered elsewhere. I'm working with whole exome seq data with the aim of identifying somatic variants. Using fastqc I can see adaptor in around 20% of reads by the end of the read (150bp PE).
I always understood that adaptor trimming was routine in this situation but I have seen a few resources recommend against it, without providing any explanation as to why. Can anyone explain why it might be detrimental to trim adaptors prior to alignment?